Increasing concern about the health risk due to solid aerosols from engine combustion has provoked more stringent imission limits, for soot particles in the range of pulmonary intrusion, at critical work-places, e.g. tunnel sites.
The scope of the joint European project VERT was restricted to the workplace "tunnel construction" and focused on the exhaust gas emission from Diesel engines in tunnels. The primary or even exclusive motivation was occupational health. Environmental aspects were a fringe motivation. The project was triggered by the large tunnel projects in Switzerland (Alpine transit), in Germany (new railway lines) and in Austria. A further impulse was the legislation, which did not exist before 1993/94, for limits of Diesel particulate emissions in all three countries.
Within the scope of the project VERT, these emissions were characterized and their effective curtailment through exhaust gas after-treatment investigated. Diesel engines, irrespective of design and operating point, emit solid particulates in the range of 100 nm, at concentrations above 10 million particulates per cm3. Specific tests indicate that a drastic curtailment of pulmonary intruding particulates is not feasible by further development of the engine combustion, nor by reformulation of fuels, nor by deployment of oxidation catalytic converters.
Particulate traps, however, can curtail the total solid particulate count, in the fine particulate range below 50 nm, by more than two orders of magnitude. The effect can be reinforced through fuel additives, such as those promoting regeneration in particulate traps.
A field test during 18 months proved that several particulate trap systems meet this requirement.
Thus the technical feasibility is basically established. The deployment of such particulate trap systems can therefore be demanded under critical exposures. Germany and Austria have already mandated the deployment of particulate traps in enclosed or partly enclosed working areas. Switzerland is preparing similar legislation.
For the study of pulmonary intruding particulates from Diesel engines and the evaluation of the particulate traps systems, the usual gravimetric evaluation of the total particulate mass PM must be abandoned or at least enhanced. This is necessary as gravimetry is non-specific with respect to the chemical composition and the aerosol properties (such as size and surface) and, hence, delivers no toxically relevant information. The counted particulate concentration appears to be the significantly more specific and sensitive criterion.
Industrially suitable and field deployable procedures should now be developed based on the dependable laboratory methods.
This paper describes the most important results relevant to the objectives of the VERT project. A more complete description of the VERT project (toxicological criteria, measurement techniques, test engines, test objects, etc.) is included in the VERT Report W11/12/97. The report is available from TTM, tel. +41.56.496.6414, fax +41.56.496.6415.
Oxidation catalytic converters are widely deployed nowadays in Diesel engines. The reason is the curtailment of particulate emissions, particularly the soluble component in automobile engines.
The conditions in automobile engines are however different from those in utility vehicle engines. Their soluble fraction of the particulates mass is a smaller percent. Hence, the basic question is whether the oxidation catalytic converters are any progress in the curtailment of solid particulate emissions.
There are different effects for the oxygen rich Diesel exhaust gas in the oxidation catalytic converter:
CO => CO2
HC => CO2 + H2O
The utility and construction site engines are mainly direct injection engines. The CO and HC emissions of such Diesel engines already have very low CO and HC emissions even without exhaust gas after-treatment. These are comparably as low as spark-ignition engines having three way catalytic converters. Hence, there is little motivation for additional efforts.
NO => NO2
SO2 => SO3
These reactions are very aggravating. NO2 is more toxic than NO. SO3 contributes to formation of sulfate particulates and to aerosols of sulfuric acid.
The AUVA measurements have shown that the oxidation catalytic converter has the effect of converting up to 40% of the NO to NO2. The ratio of NO/NO2 is usually 5 - 10% in Diesel exhaust gas. Thus the catalytic converter has substantially increased the NOx toxicity. Emissions measurements in Swedish mines have substantially confirmed this effect.
The VERT experience is shown in the figure below. The gravimetric evaluation, of exhaust gas from utility vehicular engines, indicates no curtailment in the particulates mass. Indeed a large increase is often observed.
Fig. 1: Increase in particulate mass with oxidation catalytic converter at full load
This finding was also confirmed elsewhere. The explanation is the formation of sulfates. This occurs in spite of using a modern fuel having a sulfur content of 400 ppm. Also used was a catalytic converter (manufacturer Degussa) having a selective state-of-the-art coating that should minimize the well-known sulfate formation above 350°C.
Investigations of the effects of oxidation catalytic converters on the nano-particulate emissions have been inconclusive and will be continued.
Other investigations (FAV, see Particulate Workshop ETH), too, generally show that the oxidation catalytic converter does not curtail the actual solid combustion aerosols. This is not surprising because the catalytic combustion of soot proceeds much slower than the chemical conversion in the gas phase, for which these catalytic converter are designed. A further weakness is that the precious metals, used in conventional oxidation catalytic converters, are unsuitable for this reaction.
The sum of these findings indicate that the oxidation catalytic converter should not be deployed for utility vehicle Diesel engines. The negative effects far outweigh the benefits.
Further, there are indications that additional solid particulates are created in the catalytic converter (Jing, Federal Measurement Agency EAM), This postulates that the same processes, forming particulates in Diesel engine combustion, are here repeated at low temperatures.
The hopes in the oxidation catalytic converter cannot be substantiated. A further decrease in particulate size through engine internal measures does not result in these particulates being residue-free burnt off in the catalytic converter. These processes are so much slower than the catalysis of gaseous substances and hence the intermediate retainment in a trap cannot be avoided.
Under environmental obligation, the mineral oil industry has made efforts to reformulate the Diesel fuel for curtailing emissions. The potential is being re-evaluated within the scope of the EU controlled Autooil program. The main parameters are the diminished sulfur content, decrease in aromatic components and increasing the Cetane index. The available data indicates that these measures can diminish the total particulate emission by about 5 - 15%.
This modest improvement is insufficient to attain the VERT objectives.
Hence, synthetic special fuels were investigated to evaluate the long-term emission curtailment potential of Diesel engines using reformulated fuel. These fuels might be candidates for fulfilling special conditions in small and particularly emission critical applications (e.g. tunnel construction).
The following comparison is particularly impressive:
|CH standard Diesel (SN EN 590 KO)||Special fuel (C14-20)|
|Flame point °C||55||120|
|Boiling begin °C||240|
|Boiling end 90% °C||360||360|
|Aromatic content %||< 30||< 0.1|
|Sulfur ppm||< 400||< 1.0|
|Cetane index||> 48||92 (Index)|
Compared to low sulfur standard Swiss Diesel fuel, this special fuel is a chemically pure paraffin fraction made available by the DEA-Mineralöl AG, Hamburg. The fuel contains neither sulfur nor bound nitrogen nor aromatics. It also has such a high Cetane index that it could be regarded as an ideal Diesel fuel. (Such a fuel would be the ultimate objective of ongoing efforts to reformulate the Diesel fuel but that can never be practically attained on the basis of mineral oil). The fuel performed very well in the engine, has a good fuel consumption and generally somewhat lower energy specific emissions. However, the attained emission improvements are, according to the following table, only about 10% of the VERT objective. This is disappointingly low.
|CH standard Diesel (SN EN 590 KO)||Special fuel (C14-20)|
|Fuel consumption g/kWh||221.2||210.7|
|Rated load kW||105.6||105.1|
The nano-particulate emissions were also investigated within the scope of this fuel comparison. This fuel has all the attributes for curtailing particulates formation according to prior postulate. Disappointingly, not the slightest improvement in diminishing nano-particulates emissions was attained.
A new hypothesis on the creation of solid aerosols in combustion was proposed (see dissertation Jing, Swiss Federal Measurement Agency EAM) during the course of the project. The measured results can be explained. They fit into the new hypothesis for the generation of soot particulates during the engine combustion of hydrocarbons.
Inference: The reformulation of Diesel fuel cannot curtail the emission of ultra-fine particulates. Hence, this route was abandoned.
Engine development, during the last decades, has impressively succeeded in reducing the emission of particulate mass. Many engines today emit not much more than 10% of the particulate mass prevalent 15 - 20 years ago. The required technical measures are generally: High supercharging, air intercooler, central nozzle position, increased number of nozzle perforations, very much higher injection pressure, lower air swirl, and a shallower piston bowl. Together these measures suppress the combustion close to the wall. The mixture is thus prepared in the combustion air (air preparation) and not at the surface of the combustion space, as was previously the case. All large agglomerates are thus suppressed: the black smoke disappears and thus the particulates mass curtailed.
The applicability of these measures to curtail the "nano-particulate emissions" was investigated. This was done on a low-emission engine certified 1996 in the USA. It had all the above mentioned constructional and process technical measures to diminish emissions.
Fig. 2: Comparison Liebherr 914TI (old) and 924TI-E (new) in 4 engine operating points
The new engine had an impressively low emission level compared to another engine of the same family developed previously. The oxides of nitrogen were halved. The fuel efficiency is considerably improved and simultaneously the rated power increased by 22%.
The nano-particulate distribution is here shown, normalized with load, as particulates count per kWh. There is no improvement in this distribution. Indeed, the new low-emission engine emits more ultra-fine particulate at all load points; at one load point 6 times more.
Similar observations were communicated from other investigations (USA, HEI 1/97 and ETH Workshop August 97). The discussions about the Diesel engine combustion process came to the following result. These engine developments cannot be expected to drastically decrease the particulate concentration in the nanometer range. An increase is indeed theoretically feasible.
Engine development is - if not generally, at least at this time - no strategy to effectively (factor 100) curtail the formation and emission of nano-particulates. Another solution must be sought.
The development of regenerating hot gas filters, for treating the engine exhaust gas, has been investigated in many places since more than 20 years. A multitude of filter media and regeneration systems are commercially available. Information was however scarce about the properties of these traps with regard to the penetration of nano-particulates. The only exception at the beginning of this project was the research of Kittelsson and Johnson.
Within the scope of the VERT project, it was attempted to investigate the properties of the most important filter media presently available commercially. The intention was to discover their attributes and obtain experience on their operational behavior. Particular focus was on the deposition behavior, increase in back pressure, influence of the raw emission and the regeneration response.
The following table is an overview of the behavior of the first 5 traps selected for comparison. The traps were operated at very different space velocities, thus handicapping comparison. The Unikat and Huss traps were additionally fitted with an oxidation catalytic converter.
|Filter media||Cer. Monolith +oxi. cat.||Cer. monolith + oxi. cat.||Knitted fiber filter||Wound fiber filter||Sintered metal filter|
|Construction volume l/kW||1.25||2.88||0.30||0.67||0.26|
|Space velocity 1/h||11'146||5'025||43'440||19'773||50'400|
|Pressure loss mbar||56||51||130||89||122|
|Filtration rate particulate count 20-200 nm%||92||90||90||90||86|
|Filtration rate soot after test filter analysis %||0.97||0.93||0.93||0.93||0.99|
|HC / HCo||0.46||0.45||0.72||0.56||0.65|
|Test filter analysis
Here too, the traps must be evaluated differently depending on the particulates definition. The pure gravimetric PM evaluation, including the separated hydrocarbons, sulfate and water, only attained separation rates of 80 - 85%. The first evaluation according to the particulate count (yet immature system without activated carbon trap) indicated values of 90% and more. The evaluation according to soot, based on a differentiated chemical analysis of the particulates measurement filter, indicated filtration values that were substantially more than those from the purely gravimetric PM evaluation. This means the soot traps are, regarding the soot filtration, much better than evident solely from the purely total particulates measurement. The reason is naturally that the particulates measurement also includes substances which pass through the trap in a gaseous state and are first deposited on the measurement filter. This is then meaningful when the particulate trap accelerates the sulfate formation as observed in two cases. But this definition is inappropriate for the toxicological evaluation of the system.
The 5 traps were investigated for their ultra-fine filtration properties with a somewhat immature system, i.e. without an activated-carbon trap. The results were very surprising. One of the monolithic traps filtered the large particulates very efficiently, however, there was a massive breakthrough of the smallest particulate sizes. However, the other ceramic monolith showed this breakthrough tendency only for substantially smaller particulate sizes, in a manner similar to the sintered metal trap. Substantially better was the fiber trap, in particular the deep-bed fiber trap. It was the only trap showing an improving performance for decreasing particulate size.
The investigations were repeated after further refinement of the ultra-fine particulates analysis and interposing the heated activated carbon trap. The following figures page show the results of three traps measured using the most modern instrumentation. The results are again compared on the basis of the penetration definition.
Here, too, there was a slight tendency for the sintered metal surface filter to poorer filtration with decreasing particulate size. The wound fiber trap did not have this clear tendency. The knitted fiber trap - a genuine deep bed filter - demonstrated an improvement in filtration rate with decreasing particulate size.
Fig. 3: Fine particulate filtration characteristic of a knitted fiber filter on a Liebherr 914TI engine at 1400 RPM, 50% load
Fig. 4: Fine particulate filtration characteristic of a wound fiber filter on a Liebherr 914TI engine at 1400 RPM, 50% load
Fig. 5: Fine particulate filtration characteristic of a sintered metal filter on a Liebherr 914TI engine at 1400 RPM, 50% load
These filters can perform:
This trap can be loaded until the back-pressure in new condition reaches 50 mbar. The loading of the trap should not exceed the limit of 200 mbar.
The following regeneration methods were employed on the test rig:
Extensive regeneration investigations were not performed during the engine tests. This is indeed a deficiency. It would have been highly desirable to verify the trap performance during regeneration. These observations must be postponed to the field tests.
Only the additives were investigated in a stage test regarding the regeneration temperature. The manufacturers' statements could be confirmed: Regeneration begins at 360°C.
The gravimetric evaluation shows a distinct decrease of the particulate formation when the additive is used. This effect is more pronounced in the clean gas, after the filter, than in the raw gas before the filter.
The following table demonstrates the effect of additives on Liebherr 924TI-E engine:
with Ce 100 ppm
with Fe 36 ppm
with Cu 50 ppm
with Ce 100 ppm
with Fe 36 ppm
with Cu 50 ppm
Only a weak effect was observed for HC and NOx, but a distinct improvement is seen for the particulates mass in all cases.
The question arose whether this positive effect is also found for fine particulates. The following figure, for an additive at different concentrations, shows the particulate size distribution with and without trap.
Fig. 6: Nano-particulate emissions using iron additive with sintered metal filter
This representation, too, shows the considerable curtailment of particulate formation. The particulates count, in the range of the soot peak, is reduced by half a magnitude. Below the soot peak, is the independent additive-ash peak at particulate sizes of 20 - 30 nm. At lower concentrations, this ultra-fine particulate's peak appears to decrease over-proportionally. If the concentration is further reduced, then a point is reached at which in spite of additive no independent ash particulates were measured. The inference is that the remaining ash particulates are completely deposited on carbon particulates.
The emission of such secondary particulates is not admissible. Hence, the use of additives in these concentrations cannot be permitted without trap.
Based on these results the additive technique must be positively assessed for these reasons:
The emission hazard of secondary particulates is eliminated by selecting the correct trap system.
However, the additive technology burdens the trap by retaining ash particulates. The back pressure will therefore continuously increase. When a limiting value is reached the trap must be cleaned of ash particulates. This cleansing is relatively simple (acidified warm water). The duration between cleansing should be at least 2000 operating hours.
Prerequisite for employing additive technology is a fully automated doping system with sufficient additive reserve on board. Such fully automatic additive systems are not yet commercially available for application to construction site engines.
Particulate traps with catalytic coating are increasingly available. These are designed to continuously light-off the deposited soot at sufficient temperature and oxygen reserve. The traps from the manufacturers Buck and Hug, employed in the VERT project, had such properties.
According to the manufacturers (refer also to SAE publication 960134 and 970479), the activation energy is diminished from about 140 kJ/mol to about 93 kJ/mol.
The regeneration begins at about 350°C. However, the reaction velocity at these temperatures is so small that only small soot loads are converted. Exhaust gas temperatures > 400°C are necessary to ensure regeneration of the soot burden expected in retro-fitting.
Based on these results, the statements could be confirmed for the Buck trap.
It is known that in garbage incinerators dioxins and furanes are formed at a temperature aperture around 400°C, given sufficient dwell time and the presence of carbon and soot. This reformation of dioxin and furane is accelerated when certain metals, particularly copper, are catalytically involved. The prerequisite chlorine is ubiquitous in sufficient quantities.
All these conditions are fulfilled in the Diesel particulate trap system:
If fuel additives also employed, then a possible catalytic promotion can occur.
This question must be clarified before recommending particulate trap technology, in particular the pertinent additives. A very comprehensive investigation was started within the scope of the VERT project.
Two typical trap technologies were investigated. These are a typical surface filter (SHW sintered metal filter) and a typical deep-bed filter (Buck knitted fiber filter). The traps do not generally increase the formation of dioxin and furanes. The additives iron and cerium, employed in the test, also decrease and not increase the emissions.
This result cannot be generalized. The use of copper additives are known to strongly increase dioxin formation. Copper additives were therefore excluded, for this and other reasons, from application during the VERT project.
Besides soot, polycyclic aromatic hydrocarbons always occur in the Diesel engine combustion. These PAH are known to be at least partially carcinogenic. These hydrocarbons are in a gaseous state at the exhaust gas temperatures that are expected at mean or even full load. Hence, it is often suggested that oxidation catalytic converters should be used to decompose these substances.
The oxidation catalytic converter was condemned according to the VERT criteria. Hence, it was interesting to observe how the soot filter technology influenced the PAH emissions.
A test was performed in which the operating points of the ISO 8178 C1 cycle were sequentially driven according to their weighting, and repeated during an entire day. The sum of all detectable PAH in the exhaust gas was measured. These include the components that are gaseous in the dilution tunnel plus those deposited on the filter soot at the conditions prevailing in the measurement range.
The surprising result is that the particulate trap technology can curtail the PAH sum, independent of fuel additives. The curtailment is particularly marked for the deep-bed fiber filter and attains average values of 90% and more.
This result leverages the benefits of the particulate trap technology. The explanation can only be as follows: The gaseous PAH is adsorbed on the soot filter in the trap thanks to its large active surface, and subsequently the PAH is converted during trap regeneration.
This is another reason why an additional catalytic converter is unnecessary.
The following know-how and consequences are derived from the comprehensive experimental work during the VERT project:
Thus, all prerequisites are fulfilled for wide scale deployment of this technology for improving the respiratory air quality at tunnel sites and therefore protect the occupational health of the employees.
Besides gaining this knowledge, the project activities also raised new questions. These should be solved with further investigations. One particular desire is to develop analytical methods that permit these ultra-fine particulates to be characterized according to their size and other properties, rapidly, if possible on-line and perhaps even under field conditions.